A careful examination of the construction of national income and product accounts provides remarkable insight into professional and popular thought about the operation of economic processes and activities. According to Simon Kuznets, the father of national income accounting in the United States, "being thus a summary and appraisal notion rather than an analytical entity, national income demands statistical measurement" (Kuznets  1951, 3). Kuznets quickly warned "further investigation reveals, however, that the clear and unequivocal character of such estimates is deceptive" (Kuznets  1951, 5). He then discussed the myriad of theoretical and practical issues that are encountered when one sets out to measure the economic activity of a nation. Modern students of the economy would benefit greatly from studying this article by Kuznets to understand the preconceptions that provide the foundation of the seemingly objective measurements of economic activity.
As Nelson Peach expressed it, "the kinds of activities included in National Income depend on value judgments of what is and what is not productive. Traditions, customs and mores of the community shape these judgments. Items that are included in it are different from country to country and they change over time" (Peach 1964). This statement caused me to ponder the meaning and relevance of national income accounting for the intervening forty years.
It is my objective to examine the development of a few of the major accounts and selected components of these accounts to reveal some of the judgments that were important in creating our view of economic relations. Our habitual assumptions shaped the accounts, which in turn tended to solidify our assumptions. My purpose is not to discourage quantification or the use of these accounts. In fact, institutional economists were pioneers in developing national income accounting. My point is to encourage our colleagues to reinvestigate the assumptions that provide the foundation of national economic accounts so they do not get lost in the mist of history. Furthermore, we will show that each accounting method was developed to assist in understanding a particular problem, such as social unrest, depression or war. The purpose has to be appreciated before we can understand the measuring tool. Perhaps this generation of institutional economists could develop new quantitative accounts that would provide insight into today's economic problems. At the very least, we could develop adjustments to existing accounts to correct for mistaken or out of date assumptions. According to Peach, "generally only current, productive, legal, economic activity that goes through the market is included" (Peach 1964), thus our meaning of national income is constantly changing.
The structure of our economy is also changing and this requires either adjustments to the structure of the accounts or care in interpreting the data. One simple illustration out of many cases provided by Kuznets ( 1951, 32-33) is the effect of the diminishment of the proprietor form of business organization and the growth of the corporate enterprise. Along with this transformation, employment in agriculture was in decline as manufacturing employment increased. During the period observed by Kuznets, labor's share of income was stable. But that hides the fact that many small businessmen and farmers were removed from the proprietor class and became wage earners instead. Hence, wage earners accounted for a larger portion of the productive population but, as a class, they received a constant slice of the income.
In the next section, I will discuss the historical context and purpose that drove the effort to create the measure of National Income. That is followed by a discussion of the reasons to shift from National Income to Gross National Product. (1) Next is a section that examines the introduction of the income and product accounts into modern introductory textbooks. These sections will set-up a discussion of the controversies and difficulties in developing a proper measurement of saving. Finally, there is a concluding discussion.
The Purpose for Measuring National Income
There would be little need for measuring national income or product if the worldview of the economy were that production is determined by sovereign consumers in smoothly working, competitive markets. Furthermore, these same market forces would determine the distribution of income among the owners of the factors of production. In such an economy, production and employment of labor and capital would always be maximized and the income generated by that production would be distributed efficiently and equitably by impersonal market forces to the owners of the factors of production. Any intervention by non-market institutions or agents would necessarily lead to sub-maximum and sub-optimal economic performance. If all of the axioms that are necessary to describe this system are stated, the relationships in the system can be described in mathematics without quantification.
According to Yuval Yonay (1998), institutional economists vigorously and effectively challenged the worldview of neoclassical economics during the first half of the twentieth century. Neoclassical economists were satisfied to describe the ideal economy with sophisticated models. But the models were not models of any observable economy; instead they were deductions derived from the axioms of an ideal economy (Yonay 1998, 60-65).
Having observed real problems of social strife and economic instability, several institutional and heterodox economists set out to describe the economic system through quantification. The original impetus to develop a comprehensive measure of national income in the United States was the surprisingly large block of votes cast for the Socialist Party in 1912. This coincided with the populist movement of the last quarter of the nineteenth century and the progressive movement of the first two decades of the twentieth century. Obviously many elements of the public were not convinced of the efficacy of laissez faire, and they were demanding and winning reforms.
By 1915, Scott Nearing, a socialist economist at the University of Pennsylvania had constructed estimates of the distribution of income between property owners and labor (Nearing 1915; Peach 1950, 1-2; 1964). Nearing relied on Census and Internal Revenue Service data and ended with a plea for official government data so that better estimates of income distribution could be made on a consistent basis. The question of distribution challenges a key proposition of neoclassical economics and required statistical investigation rather than deduction.
Executives at AT&T began to investigate Nearing's work. Concerned with the rise of sentiment for socialism, AT&T garnered support from like-minded corporations to secure financial support from several foundations to create the National Bureau of Economic Research (NBER) and recruited Wesley Mitchell to head the group (Peach 1964). Malcolm Rutherford (2005) thoroughly describes the complex and changing relations between the foundations and the NBER staff. An important objective of the NBER was to get a scientific measurement of the distribution of income between capital and labor.
The distribution of income was pushed down, or even off, the agenda during the conservative 1920s. However, the stability of the economic system became a front burner issue as industrialization matured and national markets displaced local markets. The reforms of the populist and progressive movements did not stymie the growth and influence of the national industrial corporation. However, these firms changed the nature of the economy and they required a stable economy. The issues of distribution and stability are not unrelated. For example, the Federal Reserve System was created in 1913 to eliminate financial panics at about the same time as the creation of the NBER.
The problem of economic stability also challenged the notion of the smoothly adjusting market system as described in neoclassical theory that always tended toward maximum production and employment. Why measure the amount of production and income if they are always maximized or quickly return to the maximum level? Wesley Mitchell, in his work at the NBER, was the pioneer of the study of business cycles. He was guided by the theoretical work of Thorstein Veblen who raised the question of the consequences of the domination of the rising industrial system by pecuniary logic and institutions (Mitchell  1947, 42, 62, 65-6). Simon...